• fourier basis in orthogonal, but not orthonormal
  • many natural and man-made phenomena exhibit an oscillatory behavior
    • after a certain amount of time, called the period, these phenomena comes back to the same position
    • conversely, we can define the frequency, i.e. the inverse of the period, that indicates the number of oscillations of the system per second
    • it thus make sense to use sines and cosines as basic building blocks to represent these oscillatory signals
    • this is the basic goal of Fourier analysis: to decompose a signal in terms of sines and cosines
  • two kinds of fourier tools,
    • fourier analysis:
      • allows determining the frequency components present in a signal
      • We move from the time to the frequency domain
    • fourier synthesis:
      • allows constructing signals with known frequency components
      • we move from the frequency domain to the time domain
  • We have started our exploration of fourier analysis with the simplest tool
  • the discrete Fourier transform (DFT) that applies to finite-length signals
  • if N is the length of the signals, we have seen that the set of complex exponential
    • \(w_k[n]=e^{j\frac{2\pi}{N}nk}\)
      • \( n=0,…,˙N−1,k=0,…,N−1 \)
    • with fundamental frequency indexed by k,
      • \( \frac{2\pi}{N}k N 2πk, \)
  • forms an orthogonal basis of \( \mathbb{ℂ}^N \). The DFT is just a change of basis in this space.